5134
Journal of the American Chemical Society
I. Kapovits and A. Kalman, Chem. Commun., 649 (1971). T. Numata and S. Oae, Tetrahedron, 32,2699 (1976), and references cited therein. J. C.Martin and M. M. Chau, J. Am. Chem. Soc., 96,3319 (1974). N. E. Waite, J. C. Tebby, R. S. Ward, M. A. Shaw, and D. H. Williams, J. Chem. SOC.C, 1620 (1971); M. J. Berenguer, J. Castells, R. M. Galard, and M. Moreno-Manas, Tetrahedron Lett., 495 (1971). M. M. Crutchfield, C. H. Dungan, J. H. Letcher, V. Mark, and J. R. Van Wazer, Top. Phosphorus Chem., 5, 1 (1967). F. Ramirez, J. A. Bigler, and C. P.Smith, Tetrahedron, 24, 5041 (1968). J. C. Martin and E. F. Perozzi, J. Am. Chem. SOC.,96, 3155 (1974). H. Budzikiewicz, C. Djerassi, and D. H. Williams, “Mass Spectrometry of Organic Compounds”, Holden-Day, San Francisco, Calif., 1967, p 222. I. Granoth, Top. Phosphorus Chem., 8, 41 (1976). and references cited therein. A. N. Hughs and M. Woods, Tetrahedron, 23, 2973 (1967).
/
100:16
/
August 2, 1978
(33) I. Granoth, J. Chem. SOC.,Perkin Trans. 2, 1503 (1972). (34) A detailed account on the crystal and molecular structure of 4, by A. V. Rivera and G. M. Sheldrick, will be published elsewhere. We thank Dr. G. M. Sheldrick for permission to disclose some of his results prior to pubiication. (35) J. A. Deiters, J. C. Gallucci, T. E. Clark, and R. R. Holmes, J. Am. Chem. Soc., 99, 5461 (1977). and references cited therein. (36) R. Sarma, F. Ramirez, B. McKeever, J. F. Marecek, and S. Lee, J. Am. Chem. SOC.,98, 581 (1976). (37) J. C.Howard, D. R. Russell, and S.Trippett, J. Chem. Soc., Chem. Commun., 856 (1973). (38) J. I. Musher, Angew. Chem., Int. Ed. Engl., 8, 54 (1969). (39) T. M. Frunze, V. V. Korshak, and V. V. Kurashev, Chem. Abstr., 54, 19553f (1960). (40) F. G. Mann and E. J. Chaplin, J. Chem. SOC.,527 (1937).
A Study of Multiple Ring Expansion and Ring Contraction Rearrangements in Observable Cycloal kyl Cat ions Roger P. Kirchen, Ted S. Sorensen,* and Kim E. Wagstaff Contribution from the Department of Chemistry, University of Calgary, Calgary, Alberta, Canada T2N 1N 4 . Received December 30. I977
Abstract: Tertiary cycloalkyl cations involving ring sizes n = 4 (small ring), n = 5-7 (common ring), n = 8-1 I (medium ring), and n = 12-20 (macroring) have been prepared. These cations have been found to undergo a very general ring expansion or contraction reaction, often in a multiple or repetitive sense. These rearrangements are facile at temperatures varying between -70 and - 100 O C and have activation barriers from 12 to 15.5 kcal/mol. Two regions of thermodynamic stability were found, the cyclohexyl ring and macrorings where n > 14 carbons. The cyclodecyl ring system sits on the energy watershed with some molecules contracting (40%) and others expanding (60%). The cyclododecyl ring system was found to be an excellent starting point for expansions to very large ring cations, involving the incorporation of up to eight carbon atoms, and some potential exists for the use of this procedure in the synthesis of large rings.
Carbocyclic ring expansion or contraction processes are a common feature of organic chemistry.’ Many of these rearrangements occur by way of carbocation intermediates and are therefore of interest in “observable carbocation studies” where the kinetics and thermodynamics of the actual expansion or contraction step can be directly probed. There have been a number of observable ion studies involving ring interconversions among alkyl-substituted cycloallyl cations (C,H2,-3+)2 but surprisingly few investigations involving the corresponding cycloalkyl cations (C,H~,-I+). Particularly attractive targets for further study are the 1n-alkylcycloalkyl cations 2, since a plausible mechanism for interconverting these rings has been known for a number of years. In 1968, two groups4 reported that the tert-amyl cation 1 undergoes a three-step degenerate rearrangement (eq 1): (1)
1,2-hydride shift to give a secondary ion, (2) 1,2-alkyl shift within the secondary ion manifold, and (3) 1,2-hydride shift to give back the tertiary ion (overall AC* = 14.8 kcal/mol). If one now applies this sequence to tertiary cycloalkyl cations 2 (eq l ) , one expects to find a general ring contraction-ring expansion process, where step 2 involves the actual expansion or contraction. However, even more significant, the mechanism can in principle be repeated and the ring might therefore continue stepwise to either contract or expand until a thermodynamically stable ring size is attained. A good way of mentally visualizing such multiple rearrangements would be to think of the cycloalkyl ring as a molecular CHrchain lasso, with the C+ center, together with the three-step eq 1 mechanism, functioning as a sort of “slip-knot’’ (see Figure 1). The incorporation of the end methyl group of the side chain into an expanding ring would be unlikely for both
1
1‘
2
2’
0002-7863/78/ 1500-5134$01 .OO/O
0 1978 American Chemical Society
Kirchen, Sorensen, Wagstaff
/ Rearrangements in Observable Cycloalkyl Cations
5135
Table 1. 13C N M R Chemical Shift Positions in the Cationsu
cation
temp, “C
a-ring CH2
C+
cu-side chain CH2
C+CH3
C+-CH2CH3
C+(CHdnCH3 (nZ2)
Others
4+-Prb 5+-Et 6+-MeC 6+-Pr 6+-Bu‘ 6+-Pent
- 1 10 -100
340.2
60.5
47.5
-100
329.2
56.1
60.8
13.3
26.9 (P-ring C H 2 ) , 20.4, 22.2
-92
327.9
55.9
59.3
14.1
6+-Hex
-80
328.0
55.0
59.3
14.3
6+-Hept
-79
328.1
55.8
59.3
14.0
7+-Pr -105 8+-MeC -90 8-Et -100 9+-Me IO+-Md 10-EtJ IO-PrJ 1I+-Md I I+-Et -105 12+-Md 12+-Pr -106
331.6
58.1
62.1
12.9
28.0 (@-ring CHz), 32.1, 27.5, 23.0, 22.4 27.7 (6-ring CH2), 31.9, 29.8, 27.0, 23.2, 22.3 27.8 (P-ring CHz), 32.2, 29.8, 29.6, ca. 27,d 23.3, 22.2 28.7 (X 2), 25.8 (X2), 20.0
332.5 329.3
54.8 (br)‘ 59.4
51.5
331.8 (br)g
54.6
54.6
332.0 ( b r ) h
57.5
57.5
l3+-Me
-89
325.5
59.8
I3+-Et/ I4+-Me/
-86 -86
331.1 33 I .O
56.5 58.8
14+-Etk I5+-Me
-8 1 -92
331.5 327.4’
56.2 59.0
26.4 (P-CH2)
10.0
12.1
37.9,ca. 36.5 (v br),‘ 27.9 (X2), 33.2 (X2), 28.6 (X2), 27.3 (X2)
12.4 (br)g
26.6 (br) (X4)g
44.8
30.2 (br) ( X 3 ? ) , h 26.5 (br) (X6?),h 22.9 28.4 (X2), 27.6 (X2), 24.9 (X4), 24.0 (X4) 27.4 (X4), 25.0 (X4). 24.1 (X2) 28.3 (X2), 27.4 (XZ), 25.9 (X2), 25.4, 22.9 (X4)
15.5 41.5
52.5
9.8 44.7 9.7 42.8
27.7 (X2), 28.2 (X2), 25.1 (X4), 25.4 (X4)
Where useful, multiplicities have been confirmed by off-resonance decoupling experiments; I3C shifts in 6 ppm relative to external Me&; double intensity peaks indicated as (X2), etc.; br = broad. Reference 8a. C Reference 9. d Peak partially obscured. e Broad owing to conformational averaging; see ref 9. f Reference 10. g Most of the peaks in this cation are broad owing to the existence of at least two conformers, interconverting relatively slowly on the N M R time scale; see ref 10. At low temperatures, one sees distinct line broadening, indicative of slow conformational averaging of I3C sites. J See Figure 2 for the actual spectrum. This cation was produced from the 15+-Me cation, forming an equilibrium mixture 15+-Me/l4+-Et = 2.0. Only the peaks characteristic of 14+-Et are quoted. In 13C-methyl labeled 15+-Me, the 13C+-13CH3coupling constant is 31.1 Hz.
’
Table 11. I3C N M R Shifts for Large Ring Expansion Seriesu 12+-Pentb 331.0 13.9 12+-Hex 331.1 14.4 12+-Hept 33 1 .O 14.5 12+-0ct 331.0 14.5 12+-Nonyl 330.9
-
-
-
I3+-BuC 326.0 13.6 13+-Pent 326.0 13+-Hex 325.9 13+-Hept 325.3 13+-0ct 324.7
-
-
-
14+-Pr 329.1 13.6 14+-Bu 328.6
ri
14+-Pent 328.3
s
14+-Hex 328.1
e
14+-Hept 328,0
2
15+-Et 329.1 9.7 15+-Prd 327.3 13.5 15+-Bu 326.5
ri
15+-Pent 326.0 or 327.3
s
s s
16+-Me 329.1 42.9 16+-Et 330.2 9.7 16+-Pr 328.0 16+-Bu 327.3 or 326.0
s
s s
17+-Me 327.7 42.2 17+-Et 330.3 9.7 17+-Pr
...
F!
F!
18+-Me 330.1 43.2 18+-Et 330.1 or 329.7 9.7
s
19+-Me 329.7 or 330.1 42.8 20+-Me
................................................................ . . .
a Symbols, etc., as in Table I. The larger number refers to the C + chemical shift and the smaller number to the CH3 carbon. Assignments of the C + chemical shifts are based partly on the sequential appearance of peaks and should be regarded as tentative in a few cases. Additional peaks can be identified, e g , a-ring CH2 carbons of the largest ring. The O-ring and cu-side chain carbons are readily identified in all of the cyclododecyl cations, occurring about 57.2 and 56.0 ppm, respectively. The higher field region consists of a fairly distinctive pattern of peaks which is similar in the various cations (see Figure 4 for an example). At least oneof the side chain carbons has an unusually low field chemical shift, ca. 37.5 ppm. Characteristically, the 12+-cations have the lowest field C + shift and the 13+-cations the highest. Also prepared from the 15-Pr alcohol.
thermodynamic and kinetic reasons5 and this entity could therefore be likened to an “end knot”. One could therefore limit (or control) expansion processes by simply limiting the
length of the side chain “rope”. For contraction processes, one can start with a simple methyl substituent and it should be noted that one would have no control in this case over the
Journal of the American Chemical Society / 100:16 / August 2, 1978
5136
@
CH, \C%/
t
Figure 1. Schematic for general ring expansion and contraction processes in observable cycloalkyl carbocations. The C+ center functions as a mechanistic “slip-knot” to allow the ring to contract or expand to the most stable size. The methyl group functions as an “end-knot’’ if the ring expansion proceeds this far.
Number 01 Carbons
In Ring
Figure 3. Schematic summary of the ring contraction and ring expansion reactions observed in this work.
Figure 2. Lower. noise-decoupled I3C N M R spectrum of the 13+-Etcation. Upper. Noise-decoupled ‘3C N M R spectrum of the resulting rearrangement product, the 14+-Me cation.
number of contraction steps since these would simply involve the extrusion of a longer and longer side chain. W e therefore undertook to prepare tertiary cycloalkyl cations from Cq (cyclobutyl) to Clo (cycloeicosyl), with varying linear alkyl side chain substituents (depending upon the particular circumstances), with a view to determining the feasibility and/or generality of these proposed stepwise contraction-expansion processes depicted in eq 1. When this study commenced, it was by no means certain that one could even prepare the larger ring tertiary cycloalkyl cations,6 let alone interconvert them. It soon became evident, however, that both these aspects were practical and we therefore extended our studies to the measurement of the kinetics involved in the expansion or contraction processes.
Results Starting with either the tertiary cycloalkyl alcohol or chloride, depending on the particular case, one can prepare and directly observe all of the initially expected ions (Le., those from
simple solvolysis), if the ion preparation is conducted at sufficiently low temperatures (-120 to -130 OC for the most unstable examples). For ring systems of greater than six carbons, ” N M R spectra are virtually featureless so that we have had to rely entirely on I3C N M R spectroscopy both to characterize all of the initially formed ions and to detect and characterize the rearrangement products (and thus also for kinetic measurements). The structures of all of the rearrangement ions were verified by direct comparison to spectra of the authentic cation, except in some cases where the ring size was 16 carbons or greater. Even in these cases, the unique position of certain I3C lines makes the assignments definitive. The I3Cpeak positions of all of the ions involved in this study are collected together in Tables I and 11. A simple illustrative example of our approach is shown in Figure 2, where the 1-ethylcyclotridecyl (henceforth abbreviated as 13+-Et) cation (produced directly from the alcohol) expands to 14+-Me (purposely constrained to one step). This 14+-Me cation spectrum shows the expected pattern of six double I3C peaks and three single^,^ with the CH3 being unique for a +C-CH3 type cation. However, this assignment is then verified by comparing the spectrum to that of the 14+-Me cation, derived directly from the corresponding alcohol. For kinetic purposes, there are several peaks unique to each ion, e.g., wring CH2, a-side-chain CH2, and CH3 of ethyl for the 13+-Et cation and wring CH2 and CH3 peaks for the 14+-Me cation. A number of the tertiary cycloalkyl cations show interesting properties in themselves; e.g., the cyclobutyl system is unique,8 the cyclohexyl cations have a twist-boat conformation which is more stable than the chair conf~rmation,~ the cyclooctyl and cyclodecyl cations show N M R line broadening due to degenerate conformational averaging9 and the cyclodecyl system shows other peculiarities possibly attributable to transannular interactions.I0 These features have been commented on and only those aspects which relate to the rearrangements will be discussed here. The Rearrangements.ThermodynamicAspects. A schematic summary of the various contraction and expansion processes observed in this work is shown in Figure 3. The corresponding kinetic data are collected together in Table 111. For rate comparison purposes, we shall be primarily interested in the AG* activation barrier column. The data in Table I11 exemplify the fact that all of the proposed contraction and expansion reactions can indeed be made to occur in a multiple rearrangement sense, in most cases quite cleanly. In retrospect, however, we have been very fortunate because there turns out to be very little temperature margin of safety. Thus the slowest rearrangements proceed at a reasonable rate at about -70 O C , but by about -50 O C , most of the larger cations (solutions) show only “baseline-noise” type
Kirchen, Sorensen, Wagstaff
/ Rearrangements in Observable Cycloalkyl Cations
5137
Table 111. Kinetic Data for the Cation Rearrangements AG*
rearrangement
---
temp, K
+
4+-Pr 5+-Et 6+-Me 5+-Et 6+-Me 7+-Pr 6+-Bu 8+-Et 6+-Bu 8+-Me 6+-Pr 9+-Me 8+-Et 7+-Pr 6+-Bu 10+-Me 6+-Pent 10+-Et s 1l+-Me k 1 ]+-Me s 10+-Et 6+-Hex IO+-Pr 12+-Me 6+-Hept 1 ]+-Et 12+-Me + IO+-Pr 12+-Pr 13+-Et 13+-Et 14+-Me 14+-Et + 15+-Me I2+-R expansions +
-
---
+
+
-+
methodo
(k0.2 kcal/ mol)
k , s-'
1.4x 10-3 8.4 x 10-4
12.3 15.4
188 168 211.5
A A A A AC AC A
8X 1.3 x 3x 7x 2x
13.3 14.3 13.9
188
B
4x I .4 x 2x 1.5 x 5.6 x 2.5 x I x
174 214 185.5
201.5
188
B
I97
A
172
B
190 20 1 197.5
AC M A
B
10-3 10-4 10-4 10-3 10-4 ( K =
I 0-4 10-3 10-4 10-4 10-3 10-3 ( K = 2.0)
b
C C
14.8
d e
13.8
.f
14.1 13.9 12.9
f
12.1
2.0)
comments
13.8 14.0 14.1 -14-15
g
h
j k
1
See Experimental Section for details. Estimated error limits in k have only been incorporated into the AG* values. Ratio of Sf-Et/6+-Me is very variable and seems to be related to the temperature used in the rearrangement. There is a very small concentration of the seven-membered cation present during the actual rearrangements. Small concentrationsof isoalkyl side chain cations are also formed in a competitive irreversible process (see ref 12). The approximate ratio of products is 8+:7+:6+= 38:47:15. e Note that this rearrangement could easily be stepwise since 9+ step is rate determining. /The rate quoted is the forward rate constant, 10+-Et I I+-Me, employing a computer simulation the I O + to obtain the three rate constants for a competitive A F! B C process. The I O + s 1 I + equilibrium ( K = 2.0) is reached well before the subsequent irreversible conversion to the 6+-cation is completed. g The product ratio is 12+:6+= 3:2. The product ratio is ca. 12+:10+= 3 : l . There is a remarkable difference in stability between the 1 1 +-Me and 1 ]+-Et cations, relative to t h e IO+-Etand IO+-Pr cations. I n the former comparison, a n equilibrium, 1 I+-Me/lO+-Et = 2.0. is observed, while in the latter, 1 I+-Et/lO+-Pr, there is no sign of an equilibrium and K must be less than 0.05. At higher temperatures, the 10+-Prcation will, of course, further rearrange to give more I2+-Me cation and some 6+-Hept cation. Overall for 1 I+-Et, expansion is preferred to contraction by a 9:l margin. J See Figure 2. Measured using labeled 15+-Me cation (see Experimental Section), employing reversible first-order kinetics. Compiex consecutive reactions which were not analyzed in detail (see Figure 4). The 12+ 13+ transformation is marginally faster t h a n the subsequent expansions in agreement with the 12+-Pr results in this
-.
table.
-
-
-
spectra indicative of a myriad of rearrangement, cleavage, or oxidative decomposition products.' I This disappearance of the cycloalkyl cations may be autocatalytic since it seems to occur quite suddenly as one raises the temperature. Additionally, in the cyclooctyl cation contraction (and apparently only here) one sees a minor product indicative of a competing secondary-primary cation contraction process.12 In a thermodynamic sense (Figure 3), there are two possible fates for the cations (given sufficient side chain carbons); they either contract or expand to form cyclohexyl cations or expand to very large ring cations (2 15-ring). The former result is unexceptional and is expected from ring stability considerations and the previous experimental work of Olah;3 however, the latter result raises some interesting possibilities which will be discussed later. These two regions of stability are not connected in a kinetic or thermodynamic sense since the previously mentioned decomposition occurs a t a much lower temperature than that needed to cause the ions to climb over the energy divide separating these regions (from either side). The division between contraction and expansion occurs in the ten-ring cations and can be nicely demonstrated with the IO+-Pr cation. About 40% of this cation contracts to the 6+-Hept cation and 60% expands to the 12+-Me cation, these processes being irreversible under our operating conditions. The 10-ring and 1 I-ring cations can be forced to only contract by providing a methyl side chain substituent but the 1 I-ring clearly prefers to expand, given the choice. The 12-ring will only expand. The relative ordering of the cation energies shown in Figure 3 can be directly determined for all expansion processes since, as previously explained, these can be made to stop after one expansion, i.e., using an ethyl side chain. Thus, 4+ 5 + , 5 + - 6 + , 1 1 + - 12+,12+-13+,and13+-.14+.Inthecontractions, we can be certain that 7+ 6+ and that 9+ 8+ but there is no direct observation for the 8+ 7+ process since kinetically the eight-ring is more stable. The 10+ 9+ ( ? )
-- -
situation is similarly unclear, also for kinetic reasons (see Table 111). The 10+ F! 1 I + cations and all cations with rings of 14 carbons or more exist in an equilibrium situation and one can, of course, directly determine their free-energy differences. These energy differences are dependent on the particular side chain and are discussed in the next section. The solvolytic behavior of ring systems from 4 to 20 carbons was originally investigated by PrelogI3 and by Brown. l 4 Differences in rate are mainly attributable to I-strain factors in the intermediate carbocations. The fact that one can now prepare all of these cations and observe directly the various ring interconversions opens the way for the direct thermodynamic measurements of the corresponding heats of reaction using scanning calorimetry. For example, the methylcyclobutyl cation, in both solvolysis14and stable ion work,8 shows signs of being particularly unstable. A AH measurement of the 4+-Et 5+-Me rearrangement and comparison to the known thermochemistry of the hydrocarbons should quantify this situation. Similar enthalpy measurements would also allow one to uncover the situation existing in 8+ 7+ (?) and I O + 9+ (?), i.e., in the former case by measuring and comparing the AH change for 8+ 6' and I+ 6'. Effect of the Side Chain on the Cation Stabilities. The Potential for a General Ring Expansion Synthesis for Very Large Ring Cations. Cycloalkyl cations up to n = 15 are obtainable from commercially available ketones. To go to larger rings, we decided to test the practicality of expanding 12-ring cations containing long alkyl side chains, making use of our lasso analogy, and the ready availability of cyclododecanone. Our suspicions were that we would obtain rather complex mixtures of ions which would be difficult to characterize. We were therefore rather surprised to find that the expansion reaction proceeds, in all of the cases studied, to give dominantly the largest possible ring, i.e., with a methyl substituent. Thus, the 12+-Pent cation first gives mainly the 13+-Bu cation and then
-
-
-
-
-
5138
Journal of the American Chemical Society
/
100:16
/ August 2, I978
since both the C+-I3CH3 and C+-CH2-’3CH3 methyl groups occupy unique positions in the I3C spectrum (see Table I). However, the larger the ultimate n ring, the more n - 1, n 2, n - 3, etc., ions that are possible and, for example, in the 18+-Me case, one expects to have a t equilibrium the dominant 18+-Me cation, the 17+-Et cation (uniquely determined), plus 16+-Pr, 15+-Bu, and perhaps 14+-Pent cations. The terminal methyl of the side chain in these latter three ions occupies a nearly identical position (see Figure 4 for an inset spectrum obtained a t a time when appreciable concentrations of these “intermediate cations” are present and note the partial resolution found for the terminal CH3 group). However, these methyls (in toto) occupy a unique position and the total concentration of all of these cations together can be approximately determined. Quenching the 18+-Me cation equilibrium mixture in sodium methoxide-methanol solution yields l-methyl-lmethoxycyclooctadecane (3) as the major ether product (see Experimental Section), along with minor amounts of the smaller ring ethers (eq 2). -80 oc W-CH2CH2CH2CH2CH2CH2CH3 -+ @Me
L 17:Et
J 15-BU
+ others
-78 OC
CH3OH
d,ksppm
343
331
OCH3
8(CH3
3i9
Figure 4. Lower, noise-decoupled I3C N M R spectrum of the 12+-Hept cation. Upper, noise-decoupled I3C N M R spectrum of the equilibrium mixture containing predominantly the 18+-Me cation. In the C+ region, we show various spectra intermediate between the two shown. In the long cidr h.n rt. h.,. v l rpcrinn te cnprtra are a l c n I.- - r-. -.i .n. t.r-.r r..n..i .n.-s .l .r.n.-. .-D .-.., cPvPrsl i n t e r m e d i n _shown. I
I
somewhat more slowly and cleanly generates the equilibrium mixture in which the 16+-Me cation is quite dominant. This same procedure was also used to generate the 17+-Me, 18+Me, 19+-Me, and 20+-Me cations, utilizing the appropriate 12+-ring species. In the 17+-Me case, we have also determined, as a check, that the 12+-Hex and 15+-Pr cations yield the same N M R spectrum. The spectra for the 12+-Hept-l8+-Me case are shown in Figure 4. In this example, the expansion from the 12-membered ring is consistent with a stepwise process since fortunately one can see distinct C+ peaks for all of the possible intermediate cations (although there is some doubt over the specific assignment). The positions of these C+ resonances are shown in Figure 4 and the peaks assigned to the lower rings rise and fall, in sequential spectra, in a manner consistent with stepwise reactions. In the 12+-Octyl case, considerable experimental care was needed in order to get reproducible results, since degradation reactions are a constant hazard and finally, in the 12+-Nonyl case, numerous attempts to obtain “clean” spectra were unsuccessful. One can obtain the 12+-Nonyl cation very easily and cleanly, but as the solution is warmed, peaks characteristic of degradation productsI5 appear simultaneously with peaks characteristic of the expanded rings. In fact, the expanded rings, once formed, seem relatively stable and the problem seems to be one of a competition at the initial 12+-Nonyl cation stage-either expansion to relatively stable ions or further attack on the 12+-Nonylcation (most likely on the side chain) by the strong acid. If the expansion reactions have gone cleanly, then the specific characterization of the ultimate and penultimate members of these various large ring cations is actually relatively easy
+ other ethers + other products
3
At first sight, one might conclude from these ring expansion results (Le., favoring the largest ring) that the rings were simply becoming more stable as they become larger. However, when one looks a t the concentrations of the corresponding (n I)+-Et, (n - 2)+-Pr, ( n - 3)+-Bu, etc., ions (n+-Me being the maximum size for a given number of carbons), these appear to have somewhat similar concentrations (and hence free energies).17 One can see this quite well in the 18+-Me cation case (Figure 4). We have previously observed a number of casesI8 where the stability of alkyl-substituted cations, in equilibrium situations, follows the hyperconjugative order, Me > Et (or n-alkyl) > i-Pr > t-Bu, and the present case seems to be (at least in part) yet another verification of this. We conclude therefore that for ring interconversions in the 4-14 carbon cations,I9 the inherent ring strain effects are the dominant thermodynamic factor, i.e., the particular alkyl substituent is relatively unimportant. However, with rings from 15 to 20 carbons, the alkyl substituent appears to exert the dominant influence. Kinetic Results. Activation Energy Differences. For the 5+ 6+ case and for rearrangements of the large rings, e.g., 13+ 14+, 14+ s 15+, etc., the activation energy barriers, AG*, are similar and in fact very similar to the barrier observed by Saunders and Hagen4 for the tert-amyl case (see Table 111). Thus, the barrier of ca. 14-15 kcal/mol appears to be typical for relatively unstrained carbocations. As the ring strain energy increases, the activation barrier generally decreases and among the least thermodynamically stable cations are the four- and nine-ring cases, which are the least stable kinetically and rearrange a t a rate about 100 times faster than the “norm”. In most instances, the contraction and expansion reactions appear to occur in a stepwise manner consistent with the eq I mechanism. This generalization breaks down, however, in the 9+-Me contraction and in the 4+-Pr expansion, where 9+-Me 8+-Et 7+-Pr 6+-Bu under temperature conditions where the eight- and seven-membered rings are stable. Preliminary investigations20aof this mechanism in the 4+-Pr case implicate two different intermediates, one leading to the 5+-Et
--
-
+
+
Kirchen, Sorensen, Wagstaff
/ Rearrangements in Observable Cycloalkyl Cations
5139
Table IV. 13C N M R Characterization of the Alcohols and Chloridesa chemical shifts alcohol
0
solvent
C-0
CH3
4-Pr
CDC13
73.4
12.1
5-Et
CDC13
82.5
8.7
6-Pr
CFCI3
71.3
16.0
6-Pent
CFCI,
71.4
14.3
6-Hex
CFC13
71.2
14.3
6-Hept
CFC13
71.3
14.3
7-Pr
CFC13
75.2
15.0
8-Et
CFC13
74.6
7.7
9-Me 9 - M e (chloride) 1 I-Et
CDC13 CFC13 CFC13
74.3 74.7 75.6
29.2 32.1 7.5
12-Pr
CDC13
75.1
14.6
12-Pr (chloride)
CFCI3
78.2
14.2
12-Pent
CFCI3
75.0
14.4
12-Hex
CFC13
75.0
14.5
12-Hept
CFC13
75.0
14.3
35.0 (X2) 41.6
12-Octyl
CFC13
74.9
14.3
35.0 ( X 2 ) 41.6
12-Nonyl
CDC13
75.1
14.0
34.4 ( X 2 ) 40.9
13-Me
CFC13
73.0
29.6
39.8 ( X 2 )
13-Et
CFC13
74.7
7.5
14-Me
CFC13
72.9
29.8
33.8 37.5 ( X 2 ) 39.7 ( X 2 )
15-Me
CDC13
73.0
28.9
39.7 ( X 2 )
15-Pr
CFCI3
73.6
14.1
37.9 ( X 2 ) 43.2
o(
carbons
36.1 (X2) 42.8 39.0 (X2) 33.8 37.9 ( X 2 ) 45.8 38.0 ( X 2 ) 43.3 38.1 ( X 2 ) 43.5 38.0 ( X 2 ) 43.5 41.8 ( X 2 ) 47.0 34.5 36.3 ( X 2 ) 35.0 ( X 2 ) 37.1 (X2) 33.6 36.2 ( X 2 ) 34.4 ( X 2 ) 43.3 37.1 ( X 2 ) 45.4 35.1 ( X 2 ) 41.6 35.1 ( X 2 ) 41.8
others 16.6. 14.4 23.8 26.6, 22.7 (X2), 16.6 33.2, 26.6, 23.2 (XZ), 22.7 ( X 2 ) 32.5, 30.7, 26.6, 23.4, 23.2, 22.7 ( X 2 ) 32.5, 30.9, 30.0, 26.6, 23.5, 23.2, 22.7 ( X 2 ) 30.4 (X2), 23.0 (X2), 17.0 28.9 (X2), 25.6, 23.0 ( X 2 ) 27.1 (X2), 21.6 (X2), 19.7 ( X 2 ) 27.4 (X2), 21.0 (X2), 20.5 ( X 2 ) 27.8 (X2). 26.6 (X2), 25.9 (X2), 21.8 (X2) 26.4 (X2), 26.1 (X2), 22.5 (X2), 22.0 (X2), 19.6 (XZ), 16.0 26.6 (X2), 26.2, 22.9 (X2), 22.5 (X2), 20.7 (XZ), 17.5 33.3. 27.1 (X2),26.6, 23.1 (X4), 22.7 (X2), 20.2 ( X 2 ) 32.9, 30.7, 27.3 (X2), 26.6, 23.3, 23.1 (X3), 22.9 (XZ), 20.4 ( X 2 ) 32.6, 31.0, 30.1, 27.1 (X2), 26.6, 23.4, 23.1 (X3). 22.7 (X2), 20.1 ( X 2 ) 32.5, 31.0, 30.3. 30.0. 27.0 (X2). 26.6, 23.3, 23.1 (X3), 22.6 (X2), 20.1 ( X 2 ) 31.8, 30.3, 29.6 (X3), 29.3, 26.4 (X2), 26.1, 22.7, 22.5 (X2), 27.1 (X2). 19.6 ( X 2 ) 28.6 (XZ), 27.3 (X2). 26.0 (X4), 22.0 ( X 2 ) 28.6 (X2), 27.3 (XZ), 26.1 (XZ), 26.0 ( X 2 j , 21.6 ( X 2 ) 27.2, 27.0 (X2). 26.6 (X2), 24.6 ( X 2 ) , 24.2 (X2). 21.4 ( X 2 ) 27.7 (X2), 26.9 ( X 2 ) . 26.6 (X4). 26.3 (XZ), 22.8 ( X 2 j 27.6 (X2), 26.6 (X2), 26.4 (X4), 25.8 (X2), 21.6 ( X 2 ) . 15.6
In 6 ppm from tetramethylsilane.
cation and one “directly” to the 6+-Me cation. From the product ratio changes with temperature, the intermediate leading to the 6+-Me cation has the lower entropy and must also have a much higher barrier for 1,2 hydride shifts. A protonated cyclopropane intermediate fits both these criteria. Similar complications of the eq 1 mechanism have also been postulated in some acyclic tert-alkyl cation rearrangements.’Ob
not well understood. However, slow addition of the alcohol to the acid during ion preparation and care in avoiding a film of acid on the upper (uncooled) portion of the N M R tube are sensible precautions. The second complication is more serious and occurs particularly in the 12+-Nonyl case. Two peaks appear in the C+-I3CH, region and one in the C+-CH2-I3CH3 region, together with extra C+ peaks and others. These results might lead one to erroneously conclude that the 20+-Me and 19+-Et cations had been formed, excep: that the stepwise aspect is completely “wrong” i n this case. Kinetic Measurements. Two general procedures were used. (A) The Experimental Section carbocation solution in an NMR tube was allowed to rearrange for Ion Preparations. These general procedures have been d e ~ c r i b e d . ~ a time in the K M R probe at a calibrated temperature ( & I “C) and For the very large carbocations, it was necessary to reduce the total quickly cooled in a lower temperature bath, and the spectrum was weight of alcohol precursor to 100-1 25 mg. There appear to be two recorded at a much lower temperature than the kinetics temperature. types of complications which may arise in the large carbocation Generally about six points were obtained. ( B ) The kinetics and the studies. The first is characterized by the appearance of tert-butyl spectra were obtained at the same temperature, taking the time as the cation peaks (and perhaps extra “area” in the saturated CH2 region) midpoint in the spectral accumulation. These kinetics were slower and as one goes to higher temperatures. It should be emphasized that, in generally four to six points were obtained. The N M R equipment and all cases, results can be obtained without this particular degradation operating conditions have been d e ~ c r i b e d In . ~ Table I l l , the conditions and the unpredictable factors which cause this t-Bu+ formation are “A” refer to the ( A ) kinetic procedure using alcohol precursors and
Journal of the American Chemical Society
5140
1:l FS03H-SbFj in S02ClF (plus S02F2 in some low-temperature cases), "AC" refers to the same kinetic procedure using chloride precursors in S b F j - S 0 2 C I F (sometimes w i t h S02F2), "ACM" refers t o the use o f chloride precursors in 1:l FS03H-SbF5 in SO2CIF. T h e "B" conditions refer t o the (B) kinetic procedure and these all involved alcohol precursors and 1:l FSO,H-SbFS in S02CIF. T h e error l i m i t s have been estimated from our a b i l i t y to reproduce the AG* barriers ( f 0 . 2 kcal/mol). since the duplicate kinetics usually involve a slightly different temperature. Labeled 1 -methylcyclopentadecyl cation (95% 1-]3CH3) was prepared in order t o obtain I3Cspectra w i t h o u t NOE enhancement and w i t h a delay t i m e between pulses. No significant differences, compared t o the n o r m a l conditions, were observed so t h a t the 13C intensity vs. concentration assumptions in this w o r k appear safe. I n any case, carbons o f similar m u l t i p l i c i t y in b o t h ions were always used in the kinetic treatment. Preparation of Alcohol Precursors. T h e alcohols I-n-propylcyclobutanol,*' I-ethylcyclopentanol,22 l - m e t h ~ l - ,I ~- n - p r ~ p y l - ,l~- n~butyl-,9 1 - n - ~ e n t y l - .1 ~ -n-hexyl~ and 1 -n-heptylcyclohe~anol,~~ 1n-propylcycloheptanol,zj I - m e t h y l c y c l ~ n o n a n o l ,1~-ethylcyclode~~ canol and l-n-propyIcyclodecano1,'0 l-methylcycloundecanol,'4a 1 -methylcyclododecano1,26 I-n-hexylcyclododecano1,271 -methylcyc l o t r i d e ~ a n o l , land ~ ~ 1 -methylcycl~pentadecanol~~~ have been reported. I-Ethylcyclooctanol, bp 58 "C (0.2 mm), was prepared by the standard G r i g n a r d synthesis; I-ethylcycloundecanol, l o w - m e l t i n g solid, was prepared f r o m the ketone and ethylmagnesium b r o m i d e using multiple treatments because o f ketone enolization; I - e t h y l c y clotridecanol, low-melting solid, b y standard G r i g n a r d synthesis; I-methylcyclotetradecanol,m p 86 "C, f r o m the ketone and m e t h y l l i t h i u m ; 1 -methylcy~lopentadecanol-/-'~C (95%) f r o m the ketone and m e t h y l Grignard-13C; I -n-propylcyclopentadecanol,mp 68 "C, by G r i g n a r d synthesis; I-n-propyl-, mp 94 "C, I-n-pentyl-, mp 69 "C, I-n-heptql-, mp 48 "C. I-n-octyl-, mp 53 "C, and l-n-nonylcyclododecanol, m p 40 O C , were prepared b y standard G r i g n a r d syntheses, chromatographing the crude product on Florisil to separate the alcohol f r o m the aldol condensation product o f cyclododecanone, 2-[ 1'-( 1'hydroxycyclododecyl)]cyclododecanone, which was characterized by 13CN M R and mass spectrometry. I-Chloro-I-ethylcyclooctane, 1 -chloro- 1-methylcyclononane, and 1-chloro- 1-propylcyclododecane were prepared in situ9 f r o m the alcohol. All o f the alcohols were checked for p u r i t y using I3C N M R spectroscopy and the results summarized in Table I V , for those not previously reported. Quenching Experiments. T h e equilibrated 18+-Me cation solution was cooled to about - 100 "C and added to a large excess o f vigorously stirred N a O C H 3 in methanol, held at -80 "C. A f t e r warming, the methanol was swamped w i t h water a n d this aqueous solution was repeatedly extracted w i t h ether. T h e combined ether extracts were dried and the ether was removed t o yield a n oily residue. Molecular distillation at 0.2 mm and a m a x i m u m temperature of 200 "C yielded about 65% volatile material. Gas chromatography analysis on C a r bowax 20M or OV- I7 showed t w o major peak regions; each consisted o f several peaks. F r o m GLC-MS analysis, the low retention group consist o f Cl9 alkanes and alkenes.28 T h e h i g h retention group were the expected C ~ethers, O m/e 296. Their GLC trace consists o f a small peak followed b y a large peak w i t h indications o f a t r a i l i n g shoulder. T h e small peak shows large M+ 43 (propyl), M+ 57 (butyl), and M + - 7 1 ions and rather small M+,M+ - 15, and M+ 29 ions. T h e trailing edge is dominated by a very large M+ 29 (ethyl) base peak while the m a i n peak shows equally dominant M+, M+ I5 (methyl), and M+ 29 ions. T e r t i a r y cyclic ethers
-
-
-
-
-
-
a( OMe R are expected t o show large M+ - R fragments, particularly where R is ethyl or longer, i.e., the parent ion is m u c h stronger in t h e m e t h y l case. These results are therefore completely consistent w i t h the cation assignments, indicating the dominant formation of I - m e t h y l - l met hoxycyclooctadecane.
Acknowledgments. The authors wish to thank the National Research Council of Canada for generous financial support.
/
100:16
/ August 2, 1978
References and Notes (1) C. D. Gutsche and D. Redmore. "Carbocyclic Ring Expansion Reactions", Academic Press, New York, N.Y., 1968. (2) (a) N. C. Deno and J. J. Houser, J. Am. Chem. SOC.,86, 1741 (1964); (b) N. C. Den0 and R. R. Lastomirsky, ibid., 90,4058 (1968); (c) V. A. Koptyug, I. A. Shleider, I. S. Isaev, L. V. Vasil'eva. and A. I. Rezvukhin, Zh. Org. Khim.. 7, 1089 (197 1); (d) G. A. Oiah, G. Liang, and Y. K. Mo, J. Am. Chem. SOC., 94, 3544 (1972); (e) T. S. Sorensen, ibid., 91, 6398 (1969); (f) T. S. Sorensen and K . Ranganayakulu, ibid., 92, 6539 (1970); (9) T. S. Sorensen and K. Rajeswari, ibid., 93, 4222 (1971); (h) K. Rajeswari and T. S. Sorensen, ibid., 95, 1239 (1973); (i) T. S.Sorensen, Tetrahedron Lett., 529 (1976); (j) K. A. Ananthanarayan and T. S. Sorensen, Can. J. Chem., 21, 3550 (1972). (3) The cyclopentyl-cyclohexyl cation rearrangement has been reported, (a) G. A. Olah, D. P. Kelly, andR. G. Johanson, J. Am. Chem. Soc., 92,4137 (1970), as have analogous one-step transformations in bicyclic ions, e.g., (b) G. A. Olah, G. Liang, J. R. Wiseman, and J. A, Chong, ibid., 94, 4927 (1972). Secondary cycloaikanois from n = 7-12 have been added to 1 : l SbFs-FSOzH and are reported3ato give the corresponding tertiary n-alkylcyclohexyl cations, except in the n = 10 case. However, the addition temperature in these experiments was only -80 OC and most of the possible intermediate tertiary ions would have further rearranged. The actual course of these contractions is therefore unknown. Various protonated ketone expansions have also been reported. (4) (a) M. Saunders and E. L. Hagen, J. Am. Chem. SOC.,90,2436 (1968); (b) D. M. Brouwer, Red. Trav. Chim. Pays-Bas, 87, 210(1968). (5) D. M. Brouwer and H. Hogeveen, Prog. Phys. Org. Chem., 9, 179-240 (1972). (6) From the work in ref 3a, it appeared possible that the rearrangement reactions might be too rapid. (7) For obvious reasons, we have not attempted to specifically assign the various ring CH? carbons beyond the eight-membered case. This also applies to the long alkyl side chain CH2's. (8) (a) R. P. Kirchen and T. S.Sorensen, J. Am. Chem. Soc., 99, 6687 (1977); (b) G. A. Olah, C. L. Jeueil, D. P. Kelly, and R. D. Porter, /bid., 94, 146 (1972); (c) G. A. Olah, R. J. Spear, P. C. Hiberty, and W. J. Hehre, ibid., 98, 7470 (1976); (d) M. Saunders and J. Rosenfeld. ibid., 92, 2548 (1970). The "degenerate rearrangement" behavior of these cations has recently been clarified and the very unusual C+ chemical shift has been examined for possible sp3 hybridization. However, this suggestion remains speculative.B8 The fact that the 4+-Pr cation rearranges to the 5+-Et and 6+-Me cations is unlikely to shed much light on the bonding picture, other than to reaffirm the basic cyclobutane skeleton. (9) R. P. Kirchen and T. S. Sorensen, J. Am. Chem. Soc., 100, 1487 (1978). (10) R. P. Kirchen and T. S.Sorensen, Can. J. Chem., manuscript in preparation. (11) Similar results are a common feature of many observable carbocation studies, although the temperature at which this occurs is quite variable. (12) The 8+-Me cation rearranges to produce some 1-isopropylcyciohexyi cation, confirmed by preparing the "authentic" cation from 1,l-dimethyl1-cyclohexylmethanol. This cation is an equilibrating system with the dominant isomer having the charge in the exocyclic side chain position. A similar rearrangement occurs with the 8+-Et cation. (13) R. Heck and V. Preiog, Helv. Chim. Acta, 38, 1541 (1955). (14) (a) H. C. Brown and M. Borkowski, J. Am. Chem. Soc., 74, 1894 (1952); (b) H. C. Brown and G. Ham, ibid., 78, 2735 (1956). (15) Hydrocarbons in SbF5-HF and SbF5-FS03H solvents are readily cleaved at either C-C or C-H bonds5 In small cations, the various CH2carbons are probably "protected" by the proximity of the positive charge.16It is rather remarkable in fact that large ring cations like 18+-Me are not more readily cleaved. (16) T. S. Sorensen, J. Chem. SOC.,Chem. Commun., 46 (1976). (17) The reported hydrocarbon Andifferences between some of the larger rings are substantial, for example, 17 > 14 > 15 16, by 5 kcal/mol; E.L. Eliel, "Stereochemistry of Carbon Compounds", McGraw-Hill, New York, N.Y ., 1962, p 189. Our AGcation ring differences are much smaller and more orderly, although to be fair one should not directly compare AGand A H values. (16) (a) T. S. Sorensen. J. Am. Chem. Soc., 89, 3782, 3794 (1967); (b) K. Ranganayakulu and T. S. Sorensen, Can. J. Chem., 50, 3534 (1972); (c) K. A. Ananthanarayan and T. S. Sorensen, ibid., 50, 3550 (1972). (19) An exception occurs with the 10+-11+ cations (see footnotes fand h of Table Ill). (20) (a) R. Kirchen, Can. J. Chem., submitted. (b) Reference 5 , pp 234-237. (21) Y. Leroux, Bull. SOC.Chim. Fr., 359 (1968). In our work, this alcohol was prepared from the ketone and propyl Grignard reagent. (22) G. Chavanne and P. Becker, Bull. SOC. Chim. Belg., 36, 591 (1927). (23) F. K. Signaigo and P. L. Cramer, J. Am. Chem. Soc., 55, 3326 (1933). (24) H. 8.Williams and W. R. Edwards, Jr., J. Am. Chem. SOC., 69, 336 (1947). (25) R. Borsdorf and B. Olesch, J. Prakt. Chem., 36, 165 (1967). (26) J. Casanova and B. Waegeli, Bull. SOC.Chim. Fr., 1289 (1971). (27) A. A. Kraevskii, A. D. Treboganov, T. D. Marieva, G. V. Varnikova, and W. A. Preobrazhenskii, Zh. Org. Khim., 3, 441 (1967). (28) The amount of alkane is variable from run to run, varying from very small in the "best" cases to being the major quench product in the "worst". The ratio of alkenes/ethers remains relatively constant, however.
-
